1
$\begingroup$

I have a large dataset like this:

+-----+------+---------+-----------+------------------+-------+-----------------+---------+-------+
| Mat | Cust | Qty(Ag) | Net V(Ag) | Demand Hits (Ag) | Qty%  | Demand Hits (%) | Net V % | Month |
+-----+------+---------+-----------+------------------+-------+-----------------+---------+-------+
| Y   | A    |     200 |      7000 |              200 | 76.9% | 95.7%           | 76.9%   |     1 |
| X   | A    |     100 |      1000 |               10 | 83.3% | 66.7%           | 33.3%   |     1 |
| Y   | C    |      50 |      1750 |                8 | 19.2% | 3.8%            | 19.2%   |     1 |
| X   | B    |      20 |      2000 |                5 | 16.7% | 33.3%           | 66.7%   |     1 |
| Y   | B    |      10 |       350 |                1 | 3.8%  | 0.5%            | 3.8%    |     1 |
| Y   | A    |     600 |     21000 |               78 | 86.0% | 86.7%           | 86.0%   |     2 |
| X   | B    |      60 |      6000 |               56 | 0.6%  | 50.5%           | 0.6%    |     2 |
| X   | C    |   10000 |   1000000 |               45 | 98.9% | 40.5%           | 98.9%   |     2 |
| Y   | B    |      98 |      3430 |               12 | 14.0% | 13.3%           | 14.0%   |     2 |
| X   | A    |      50 |      5000 |               10 | 0.5%  | 9.0%            | 0.5%    |     2 |
+-----+------+---------+-----------+------------------+-------+-----------------+---------+-------+

This data is about sale of materials (Mat) to customers (Cust). For each Mat-Cust combination I have the qty they ordered (Qty), net value of their orders (Net V) and the number of orders (Demand Hits) and the month (Month).

I also have percentages of orders, net value, and quantity. For example: In the line one (Mat Y and Cust A) the Qty% is 76.9%. This means that 76.9% of the total quantity of material Y was ordered by customer A.

So if you sum the Qty % of Mat Y-Cust A and Mat Y - Cust C and Mat Y - Cust B for month 1 it will add up to a 100%.

The same has been done for the other percentages as well.

I was looking for a good way to visualize this data (That could run into millions of records even for one month) and also to statistically analyse it to draw any conclusions.

Any ideas for an exploratory visual and statistical analysis?

Here are some questions I am thinking about?:

  • Are there customers who make up a large portion of the sales of particular items?
  • Are there specific Mat-Cust combos that can be considered "Critical" based on demand hits or net value?
  • Can I cluster Mat-Cust into different categories and treat each category differently based on their unique characteristics?
  • Is customer buying patterns repeating across time periods?
$\endgroup$
  • 2
    $\begingroup$ What are your goals for the visualization? Do you want to see if you can discover a pattern? (What kind of pattern?) Do you want to present a finding? (What finding; to whom?) Do you want to assess an assumption or a hypothesis? (What?) Etc. $\endgroup$ – gung Jun 21 '17 at 19:54
  • $\begingroup$ @ gung : Here are some questions I am thinking about? : - Are there customers who make up a large portion of the sales of particular items? - Are there specific Mat-Cust combos that can be considered "Critical" based on demand hits or net value? - Can I cluster Mat-Cust into different categories and treat each category differently based on their unique characteristics? - Is customer buying patterns repeating across time periods? $\endgroup$ – Achyutha Mohan Jun 21 '17 at 20:14

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.